In one embodiment, a device obtains a machine learning model indicative of how to focus on particular location information from a plurality of radio frequency (RF) elements to provide an accurate location estimate of a wireless client based at least in part on angle-of-arrival information of the wireless client. When the device then obtains location information regarding the wireless client from the plurality of RF elements, it may apply the machine learning model to the location information regarding the wireless client to focus on particular location information of the location information from the plurality of RF elements. The device may then estimate a physical location of the wireless client based on focusing on the particular location information during a locationing computation.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method, comprising: obtaining, at a device, a machine learning model indicative of how to focus on particular location information from a plurality of radio frequency (RF) elements to provide a location estimate of a wireless client based at least in part on angle-of-arrival information of the wireless client; obtaining, at the device, location information regarding the wireless client from the plurality of RF elements; applying, by the device, the machine learning model to the location information regarding the wireless client to focus on particular location information of the location information from the plurality of RF elements, wherein focusing the particular location information includes: weighting location information from certain antennas of certain access points to define a level of influence of the location information from certain antennas to the locationing; and estimating, by the device, a physical location of the wireless client based on the focusing on the particular location information during a locationing computation.
2. The method as in claim 1 , wherein the location information comprises locational probability heatmaps.
3. The method as in claim 1 , wherein obtaining the machine learning model comprises: training the machine learning model based on location information regarding one or more wireless clients in known locations.
4. The method as in claim 1 , wherein focusing on the particular location information during a locationing computation comprises: excluding location information from certain access points.
5. The method as in claim 1 , wherein weighting is binary to either include or exclude particular location information.
A system and method for processing location information involves determining the relevance of location data to a user's context, such as their current activity or preferences. The method includes analyzing location data from multiple sources, such as GPS, Wi-Fi, or cellular signals, to assess its reliability and accuracy. The system then applies a weighting mechanism to the location data, where the weighting is binary—either fully including or excluding specific location information based on predefined criteria. This binary weighting ensures that only the most relevant and accurate location data is used for further processing, such as navigation, tracking, or personalized recommendations. The method may also involve filtering out noisy or inconsistent location data to improve the overall accuracy of the system. By dynamically adjusting the inclusion or exclusion of location information, the system provides more precise and context-aware location-based services. This approach is particularly useful in applications where location accuracy is critical, such as emergency response systems, autonomous vehicles, or location-based advertising. The binary weighting simplifies the decision-making process by eliminating intermediate weight values, ensuring clear and decisive filtering of location data.
6. The method as in claim 1 , wherein weighting is within a range from zero to a maximum weight value.
This invention relates to a method for assigning weights to data elements in a computational system, addressing the challenge of optimizing weight distribution to improve performance in applications such as machine learning, data analysis, or decision-making processes. The method involves determining a weight for each data element, where the weight represents its relative importance or influence in the system. The key innovation is that the weighting is constrained within a predefined range, specifically from zero to a maximum weight value. This ensures that weights do not exceed a specified upper limit, preventing excessive influence by any single data element and maintaining system stability. The method may also include adjusting the weights dynamically based on input data or system requirements, ensuring adaptability. By limiting weights to a controlled range, the invention improves reliability, efficiency, and fairness in systems where weighted data elements are processed or analyzed. This approach is particularly useful in scenarios where unconstrained weights could lead to biased outcomes or computational inefficiencies. The method may be applied in various domains, including but not limited to recommendation systems, risk assessment models, and optimization algorithms.
7. The method as in claim 1 , wherein the location information is based on angle-of-arrival information and one or more values selected from a group consisting of: received signal strength indication (RSSI), angle-of-arrival phase value, signal variation, noise floor, variance of samples, and channel condition.
A method for determining the location of a device within a wireless network involves using angle-of-arrival (AoA) information combined with additional signal metrics to improve accuracy. The technique leverages multiple signal characteristics to refine position estimates, addressing challenges in indoor or complex environments where traditional methods like GPS may fail. The AoA information provides directional data, while supplementary values such as received signal strength indication (RSSI), angle-of-arrival phase values, signal variation, noise floor, variance of samples, and channel conditions enhance precision. These metrics help mitigate multipath interference, signal fading, and other environmental factors that degrade location accuracy. By integrating these diverse measurements, the method achieves more reliable positioning, useful for applications like asset tracking, indoor navigation, and IoT device localization. The approach is particularly valuable in scenarios requiring high-precision location data without relying solely on GPS or other satellite-based systems.
8. The method as in claim 1 , wherein focusing on the particular location information during a locationing computation comprises: including location information from only certain access points; and weighting location information from certain antennas of those included certain access points to define a level of influence of the location information from certain antennas to the locationing computation.
9. The method as in claim 1 , wherein the machine learning model is based on an artificial neural network (ANN).
10. A tangible, non-transitory, computer-readable medium storing program instructions that cause a computer to execute a process comprising: obtaining a machine learning model indicative of how to focus on particular location information from a plurality of radio frequency (RF) elements to provide a location estimate of a wireless client based at least in part on angle-of-arrival information of the wireless client; obtaining location information regarding the wireless client from the plurality of RF elements; applying the machine learning model to the location information regarding the wireless client to focus on particular location information of the location information from the plurality of RF elements, wherein focusing the particular location information includes: weighting location information from certain antennas of certain access points to define a level of influence of the location information from certain antennas to the locationing; and estimating a physical location of the wireless client based on the focusing on the particular location information during a locationing computation.
This invention relates to wireless client location estimation using machine learning and radio frequency (RF) elements. The problem addressed is improving the accuracy of location estimation by selectively focusing on relevant location information from multiple RF elements, such as antennas in access points, while mitigating interference or noise from less relevant data. The system involves a machine learning model trained to determine how to prioritize location information from various RF elements based on angle-of-arrival (AoA) data of the wireless client. The model processes real-time location data from the RF elements, applying learned weights to adjust the influence of specific antennas or access points in the location estimation process. By dynamically weighting the contributions of different RF elements, the system enhances the precision of the client's physical location calculation. The machine learning model is stored on a non-transitory computer-readable medium and executed to obtain location data, apply the learned weighting scheme, and compute the client's location. This approach improves location accuracy by leveraging AoA information and selectively emphasizing or de-emphasizing data from certain antennas or access points, depending on their relevance to the current location estimation. The method is particularly useful in environments with multiple RF sources where interference or multipath effects can degrade location precision.
11. The computer-readable medium as in claim 10 , wherein focusing on the particular location information during a locationing computation comprises: excluding location information from certain access points.
12. An apparatus, comprising: one or more network interfaces to communicate with a network; a processor coupled to the network interfaces and configured to execute one or more processes; and a memory configured to store a process executable by the processor, the process, when executed, configured to: obtain a machine learning model indicative of how to focus on particular location information from a plurality of radio frequency (RF) elements to provide a location estimate of a wireless client based at least in part on angle-of-arrival information of the wireless client; obtain location information regarding the wireless client from the plurality of RF elements; apply the machine learning model to the location information regarding the wireless client to focus on particular location information of the location information from the plurality of RF elements, wherein focusing the particular location information includes: weighting location information from certain antennas of certain access points to define a level of influence of the location information from certain antennas to the locationing; and estimate a physical location of the wireless client based on the focusing on the particular location information during a locationing computation.
13. The apparatus as in claim 12 , wherein focusing on the particular location information during a locationing computation comprises excluding location information from certain access points.
14. The apparatus as in claim 12 , wherein weighting is binary to either include or exclude particular location information.
Cooperative Patent Classification codes for this invention.
July 25, 2019
October 6, 2020
Discover thousands of AI-analyzed patents with comprehensive breakdowns, multimedia content, and expert insights.